3

Many diachronic processes of language change appear to derive from synchronic errors in linguistic performance. How do generativists account for this if performance and competence are separate? If errors in linguistic performance are not considered part of "language," how do they cause changes in language?

1
  • Not all generativists believe every random thought that Chomsky has articulated. Only some.
    – jlawler
    Jun 12, 2022 at 15:07

2 Answers 2

3

Linguistic performance is part of "language", in the E-language sense. Generative theory is a theory of the computational mechanism, which underlies human speech output. So the grammar say "Do X", but a speaker does not always follow those rules, or, the rules may produce an output which is under non-optimal conditions (i.e. normal conditions) indistinguishable from a different output. The grammar is induced on the basis of the ambient data (a grammar cannot be directly transmitted from mind to mind), so to the extent that the data diverges from what the grammar would compute, the language will change. This is discussed extensively in Mark Hale's Historical Linguistics: Theory and Method.

5
  • It seems, though, that Chomsky largely rules out linguistic performance as being part of linguistic theory: "Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech-community, who knows its (the speech community's) language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of this language in actual performance."
    – ubadub
    Nov 15, 2018 at 18:26
  • 1
    Also, the process you've described seems kind of random, and yet there seem to be set patterns of linguistic change that repeat throughout history in a wide variety of language families. It seems like you need a coherent account of how I-language and E-language relate in order to explain that.
    – ubadub
    Nov 15, 2018 at 18:32
  • 1
    That's a quirk of Chomsky's use of "linguistic" to mean specifically "grammatical". And I would say that Hale provides that coherent account of how I-language relates to E-language at least from the historical POV. The appearance of randomness decreases once you consider how likely it is that a particular error will arise (for example, post-nasal voicing is highly likely to arise, post-nasal devoicing is unlikely to arise, but both do happen).
    – user6726
    Nov 15, 2018 at 18:51
  • 1
    Re: randomness, you can describe some errors systematically, no? For example, the familiar apron/napron, ekename/nickname, un os/un nos issue is easy to formulate. It's not the most interesting error but it gives an idea. Nov 15, 2018 at 19:03
  • 1
    In fact I don't know of any evidence that GG believes in true randomness. Instead, it believes that certain explanations are outside the domain of the theory of computation.
    – user6726
    Nov 15, 2018 at 19:42
1

A younger generation mistakenly takes their elders' errors to be rule-produced. So they have to make up new rules to explain what they hear. We wouldn't have this problem if the rules could be codified and taught.

7
  • If the elders language errors are consistent then they must be produced by alternative rules. Does your 'could' mean 'would',or do you mean that we use cognitive processes that are beyond conscious comprehension?
    – amI
    Nov 16, 2018 at 5:58
  • 1
    @amI, the elder generation errors may be performance errors -- why would they be consistent with each other or with any system? That's the point here. It is the next generations' task to rationalize the language they hear into a new system, derived from the earlier system with performance errors mixed in.
    – Greg Lee
    Nov 16, 2018 at 6:49
  • Oh -- but a spurious explanation of a random error is likely to be corrected before becoming infectious.
    – amI
    Nov 16, 2018 at 8:02
  • @aml, Oh, really? Perhaps because of the direct mind-to-mind contact that user6726 mentioned above?
    – Greg Lee
    Nov 16, 2018 at 14:52
  • We seem to be talking past each other -- My 'correction' is made by the listener, precisely because the error (if random) is inconsistent with subsequent exposure. If the speaker's error is instead repeated (internalized by the speaker due to a misinterpreted rule), then the listener must deduce a compatible rule. I can see lexical changes perhaps coming from single errors (since they are just data, not rules), but I think syntactic (rule) changes require repeated examples. Can performance errors fit that pattern?
    – amI
    Nov 17, 2018 at 4:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.